Hepatocyte growth factor (HGF) has been shown to enhance recovery from renal tubular ischemia. We investigated the possibility that HGF improves recovery by preventing ischemia-induced loss of cell adhesion. Murine inner medullary collecting duct-3 (mIMCD-3) cells subjected to 90% ATP depletion demonstrated a 55% decrease in adhesion, an effect that was completely reversed by the addition of HGF. Assays examining release of adherent cells revealed similar results with 30 min of ATP depletion causing loss of adhesion of 25% of mIMCD-3 cells and HGF completely reversing this effect. In contrast, HGF was unable to reverse the loss of adhesion of cells exposed to 99% ATP depletion. Examination of the mitogen-activated protein kinase (MAPK) signaling pathway revealed that HGF could induce extracellular signal-regulated kinase (ERK) phosphorylation in control and 90% ATP-depleted cells but not in 99% ATP-depleted cells. Inhibition of ERK activation with U0126 completely blocked the HGF-dependent reversal of ATP-depleted cell adhesion. Thus ATP-depleted cells demonstrate a marked decrease in cell adhesion that is reversible by the addition of HGF. This effect of HGF requires activation of the MAPK pathway.
- extracellular signal-regulated kinase
- renal ischemia
- inner medullary collecting duct
- mitogen-activated protein kinase
- adenosine 5′-triphosphate
hepatocyte growth factor(HGF) is a circulating protein that is produced primarily by fibroblasts and peritubular interstitial cells in a variety of organs including liver, kidney, lung, and pancreas (8, 31, 46,52). The addition of HGF to cultured epithelial cells may result in at least three distinct responses: mitogenesis, motogenesis (scattering/chemotaxis), and morphogenesis (branching tubule formation) (5). These responses are mediated by the c-met tyrosine kinase receptor and are likely to be involved in developmental and repairative epithelial cell migration and branching tubule formation (8, 18, 44, 46).
There is significant evidence to suggest that HGF and its receptor c-met play an important role in determining tubular cell fate after acute tubular injury. In the adult rat kidney, several laboratories have found that transient injury results in rapid upregulation of both HGF (at multiple sites) (16, 17) and c-met (selectively at the site of injury) (18, 25), suggesting a role for this growth factor receptor in tubular repair. Similarly, HGF levels were increased after tubular injury with either CCl4 or HgCl2 (21, 30). Two studies have shown accelerated recovery from models of acute tubular necrosis in rodents after the administration of HGF. In a study by Miller and colleagues (29) examining renal artery cross-clamp in the rat, the administration of HGF 30 min after reperfusion decreased the peak rise of blood urea nitrogen (BUN) and creatinine and hastened the recovery back to baseline renal function (29). By using exposure to either HgCl2 or cisplatin in mice to cause acute tubular injury, Kawaida and co-workers (21) found that pretreatment with HGF prevented the doubling of BUN that occurred in control mice at 24 and 48 h after toxin exposure. In both studies, histology revealed that control animals had widespread cell necrosis and areas of denudation of the tubular basement membrane, whereas HGF-treated animals had almost no areas of tubular cell loss. Thus HGF appears to not only enhance recovery but can actually diminish the acute injury.
After renal ischemia, the initial response of the tubular epithelial cells is loss of both polarity and tight junctional integrity followed over hours to days by one of three cell outcomes: frank necrosis, detachment and apoptosis, or dedifferentiation into a more fibroblast phenotype (thinning and spreading of the cytoplasm to cover the denuded basement membrane) (2, 47). Many of the cells that detach after acute renal injury are actually viable and can be cultured from urine (12, 26,35-37). These cells, instead of dedifferentiating and serving to aid in repopulation and recovery, are a contributor to the luminal obstruction that occurs after acute injury and therefore actually serve to worsen acute renal failure (23, 47). The mechanism by which otherwise viable cells undergo detachment from the basement membrane, and strategies to prevent this, are therefore of great interest. We investigated the possibility that part of the mechanism by which HGF can minimize renal tubular injury and promote repair is by enhancing cell adhesion of ischemic, ATP-depleted epithelial cells.
In an attempt to mimic the effects of in vivo ischemia on cell adhesion, we used a combination of the mitochondrial inhibitor oligomycin A and the glycolytic inhibitor 2-deoxyglucose to induce ATP depletion in cultured murine inner medullary collecting duct-3 (mIMCD-3) cells (28, 43). Conditions were determined whereby cultured renal epithelial cells are substantially ATP depleted for 30 min but recover if ATP levels are restored. We found that ATP-depleted cells failed to adhere to culture plates and that this diminished adhesion was reversed by treatment with HGF. In addition, we found that 25% of adherent cells would release attachment in the setting of ATP depletion, an effect that was again reversed by treatment with HGF. This response is fully prevented by the mitogen-activated protein kinase (MAPK) kinase (MEK) inhibitor U0126, suggesting that HGF-mediated cell adhesion is dependent on activation of the MAPK-signaling pathway.
MATERIALS AND METHODS
Cells and reagents.
Experiments were performed with mIMCD-3 cells, immortalized murine inner medullary collecting duct cells of ureteric bud origin (39) that have been shown to express the c-met receptor and undergo migration and tubulogenesis in response to HGF (5). All cells were grown in DMEM-Ham's F-12 medium (DMEM/F-12) media supplemented with 10% FCS in a 5% CO2environment. HGF, 2-deoxy-d-glucose, oligomycin A, and poly-l-lysine were obtained from Sigma. Laminin, fibronectin, and collagen type IV were obtained from Becton Dickinson. Type I collagen was obtained from Upstate Biotechnology, and U0126 was obtained from Promega.
ATP depletion model.
mIMCD-3 cells were bathed in human peripheral-type benzodiazepine receptor (HPBR) buffer solution [(in mM) 135 NaCl, 5 KCl, 3.33 NaH2PO4, 0.83 Na2HPO4, 1 CaCl2 , 1 MgCl2, and 5 HEPES, pH 7.4] for 20 min and then exposed to HPBR containing either 10 mM 2-deoxyglucose and 5 μM oligomycin A or 100 μM 2-deoxyglucose and 25 nM oligomycin A for 30 min (3, 28, 50). For recovery experiments, the cells were washed once with HPBR buffer, and the media were replaced with standard DMEM/F-12 media containing 17.5 mM glucose. ATP levels were determined using the Bioluminescence Detection Kit for ATP (Promega). Briefly, mIMCD-3 cells were extracted with 2.5% trichloroacetic acid, the sample was neutralized with 250 mM Tris-acetate (pH 7.75), and ATP levels were measured using the Luciferase/Luciferin (L/L) Regent according to the manufacturer's protocol. A 2-s delay time after L/L Reagent injection and a 10-s relative light unit signal integration time were used. An ATP standard curve was obtained by adding 10 μl of ATP standard diluted to the proper concentrations and assayed using L/L Regent.
Cell adhesion assay.
To examine cell adhesion on different matrix substrates, plastic tissue culture dishes were coated overnight at 4°C with either laminin (1 μg/ml in DMEM/F-12), type I collagen (1 μg/ml in PBS), type IV collagen (1 μg/ml in 0.05 N HCl), fibronectin (1 μg/ml in DMEM/F-12), or polylysine (0.01% in water) as per manufacturer's instructions. mIMCD-3 cells were harvested by bathing in 1 mM EDTA for 10 min. The cells were then counted, and 4 × 104cells/well were plated on the coated 96-well plates in DMEM/F-12 media for the indicated time. Nonadherent cells were removed by being washed once with PBS, and remaining cells were quantitated by the method of Oliver et al. (34). Briefly, cells were fixed for 30 min with 10% formalin, stained for 30 min with 1% methylene blue, and washed three times with 0.01 M boric acid (pH 8.5). The intracellular dye was extracted with 0.1 N HCl/ethanol (1:1) for 1 h at room temperature and quantified by absorbance at 655 nm.
To examine adhesion after ATP-depletion, cells were harvested with 0.5% trypsin, 5.3 mM EDTA for 5 min. Equal numbers of suspended cells were incubated in either control media (DMEM/F-12) or HPBR buffer for 20 min. Trypsin had to be utilized to harvest cells for these experiments to prevent excessive cell clumping during the incubation period. Appropriate concentrations of 2-deoxyglucose and oligomycin A were then added to the cells suspended in HPBR for an additional 30 min. The cells were then pelleted, resuspended in DMEM/F-12 ± HGF (40 ng/ml), and plated for 30 min on a 96-well plate. For inhibitor studies, 20 μM U0126 was added at the same time as the HGF. Nonadherent cells were removed by washing once with PBS, and adherent cells were quantified, as above. To compare multiple experiments, adhesion of control cells was designated as 100%, and all other values were normalized to the control value. Each individual well representsn = 1, and each experiment was repeated on at least two separate occasions. P values were obtained using the Student's t-test.
Cell release assay.
mIMCD-3 cells were harvested by bathing in 1 mM EDTA for 10 min, counted, and 4 × 104 cells/well were plated on 96-well plates in DMEM/F-12 media and allowed to adhere for 2 h. Nonadherent cells were removed by being washed once with PBS, and the media were changed to control media (DMEM/F-12) or HPBR buffer for 20 min. Cells in the HPBR buffer were then ATP depleted for 30 min by the addition of 2-deoxyglucose and oligomycin A ± HGF. Cells that had released from the plate after the 30-min incubation were removed by washing once with PBS, and remaining cells were fixed and quantitated, as described above.
Western protein analysis.
Adherent mIMCD-3 cells were serum starved overnight followed by incubation in control media or ATP depletion media for 30 min, as described. In the MEK inhibition experiments, 20 μM U0126 or vehicle control was added for 20 min during the 30-min incubation in ATP depletion buffer. The cells were then stimulated with 40 ng/ml HGF or vehicle control for 10 min, lysed with radioimmunoprecipitation assay lysis buffer [(in mM) 160 NaCl, 20 Trizma base, 1 EDTA, 1 EGTA, 1 dithiothreitol, 1% triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecylsulfate, 1 sodium orthovanadate, 1 sodium fluoride, 1 phenylmethylsulfonyl fluoride, 0.5 μg/ml leupeptin, and 0.5 μg/ml pepstatin A], and insoluble debris removed by centrifugation. Forty micrograms of cell lysates were then separated on 10% SDS-PAGE and immunoblotted with either anti-phospho MAPK or anti-extracellular signal-regulated kinase (ERK) (New England Biolabs).
HGF increases renal epithelial cell adhesion.
Initial experiments were designed to determine whether HGF can enhance adhesion of control mIMCD-3 cells. A time course of adhesion revealed that HGF had no effect on initial cell attachment at 5 min but caused a more rapid increase in adhesion at 30 and 60 min (Fig.1). By 2 h, all of the cells were adherent, with no difference between HGF-treated and control cells. We also compared adhesion at 30 min to commercial plastic tissue culture dishes coated with laminin, type I collagen, type IV collagen, or fibronectin vs. cell adhesion to poly-l-lysine-coated dishes. Consistent with the results seen in carcinoma cells (15), treatment with HGF increased mIMCD-3 cell adhesion to plastic dishes, type I collagen, and fibronectin (data not shown). In contrast, HGF had no effect on attachment to dishes coated with laminin, type IV collagen, or poly-l-lysine in these cells.
Reversible ATP depletion in mIMCD-3 cells.
In an attempt to mimic in vivo transient ischemia, we utilized the combination of the mitochondrial inhibitor oligomycin A and the glycolytic inhibitor 2-deoxyglucose to induce ATP depletion. mIMCD-3 cells exposed to 25 nM oligomycin A and 100 μM 2-deoxyglucose for 30 min demonstrated an ∼90% decrease in intracellular ATP concentration (Fig. 2 A), whereas ATP levels in cells treated with 5 μM oligomycin A and 10 mM 2-deoxyglucose for the same time period declined to 1% of baseline values (Fig.2 B). Of note, treatment with HGF did not alter the degree of ATP depletion induced by oligomycin A and 2-deoxyglucose (90% ATP depletion = 9.3 ± 0.5% of control ATP levels, 90% ATP depletion + HGF = 11.0 ± 0.58%; 99% ATP depletion = 1.82 ± 0.1%, 99% ATP depletion + HGF = 2.02 ± 0.1%, n = 3 for all conditions). When the oligomycin A and 2-deoxyglucose were removed and cells were cultured in glucose-containing media, ATP levels fully recovered within 1 h in the 25 nM oligomycin A/100 μM 2-deoxyglucose-treated (90% ATP-depleted) cells but recovered to only 20% of baseline values in the 5 μM oligomycin A/10 mM 2-deoxyglucose-treated (99% ATP-depleted) cells. After 30 min of ATP depletion, 1.3% of 90% ATP-depleted cells and 2.3% of 99% ATP-depleted cells were judged to be dead by trypan blue staining. When these ATP-depleted cells were allowed to recover for 2 h suspended in normal media, 1.7% of 90% ATP-depleted cells and 2.7% of 99% ATP-depleted cells were trypan blue positive. In contrast, 5.3% of cells maintained for 2 h in 90% ATP depletion media were trypan blue positive and 34% of cells maintained for 2 h in 99% ATP-depletion media were judged to be dead. Therefore, 30 min of ATP depletion results in cells that appear to remain viable if resuspended in normal, glucose-containing media.
ATP depletion causes reversible loss of cell adhesion.
To examine the effects of ATP deletion on cell adhesion, suspended mIMCD-3 cells were subjected to 30 min of ATP depletion media or control media before adhesion for 30 min on 96-well plates. Ninety percent ATP depletion resulted in a 55% decrease in mIMCD-3 cell adhesion compared with control cells, whereas 99% ATP depletion resulted in <1% of the cells adhering after 30 min (Fig.3 A). The addition of HGF at the time of plating again resulted in increased adhesion of control cells and completely reversed the loss of adhesion seen in 90% ATP-depleted cells. In contrast, HGF treatment of mIMCD-3 cells subjected to 99% ATP depletion resulted in only a modest increase in the number of cells adhering to the plate.
In an attempt to mimic the conditions of cell detachment after ischemic acute renal failure, we examined the effect of ATP depletion on detachment of adherent cells. mIMCD-3 cells that were plated and allowed to fully adhere (2 h) were subjected to ATP depletion for 30 min followed by the removal of detached cells and quantitation of remaining cells. As has been described in the renal tubule after transient ischemia, 90% ATP depletion resulted in the detachment of 25% of the cells at 30 min, whereas 99% ATP depletion causes ∼40% of the cells to lose attachment (Fig.3 B). Of note, in agreement with the in vivo results of Racusen (35), 45% of the cells that detached after 30 min of 90% ATP depletion reattached and proliferated when replated in normal culture media, demonstrating that these cells were viable at the time of detachment (data not shown). Similar to the effects seen in the ATP-depleted cell adhesion assay, the addition of HGF at the time of ATP depletion completely prevented the release of mIMCD-3 cells subjected to 90% ATP depletion. In 99% ATP-depleted cells, there was no significant effect of HGF on cell detachment.
We then examined the effects of ATP depletion on mIMCD-3 cell morphology. When these cells are plated on plastic dishes, they are initially rounded. In the absence of ATP depletion, over several hours the cells extend processes in multiple directions and flatten out on the plate (Fig. 4, top). In contrast, when cells were either 90% ATP depleted or 99% ATP depleted before plating, the process of flattening and of extension was severely inhibited (Fig. 4, middle left and bottom left). Treatment with HGF at the time of plating had no further effect on spreading of control cells at 2.5 h (top right) but dramatically reversed the loss of process extension and flattening observed in 90% ATP-depleted cells (middle right). In 99% ATP-depleted cells, HGF failed to reverse the loss of process extension and flattening in the majority of cells (bottom right).
The HGF-mediated increase in cell adhesion is MAPK dependent.
HGF stimulation of mIMCD-3 cells results in activation of multiple signaling pathways including the MAPK proteins ERK1 and ERK2 (9,13). We have demonstrated that activated ERK2 can bind and phosphorylate the membrane-associated protein GAB1 after HGF stimulation (41), and active ERK has been recently demonstrated to associate with focal adhesions at the sites of cell attachment after src activation (11). These results suggest that activation of ERK can result in its recruitment to focal adhesion complexes where it might phosphorylate proteins that are critical for cell adhesion. To determine whether activation of ERK is critical for the observed HGF effects on ATP-depleted cell adhesion, we first investigated whether HGF can stimulate ERK activation in ATP-depleted cells. Adherent mIMCD-3 cells were pretreated with control media or ATP depletion media ± the MEK inhibitor U0126 for 30 min, followed by 10 min of stimulation with HGF. Whole cell lysates were then probed with an antibody directed against activated ERK1/2. The addition of HGF to control mIMCD-3 cells resulted in phosphorylation and activation of ERK1 and 2 as has been previously described (Fig. 5 A). This phosphorylation of ERK after HGF stimulation still occurs in 90% ATP-depleted cells but fails to occur in 99% ATP-depleted cells. Pretreatment with the MEK inhibitor U0126 completely prevents ERK activation. Reprobing of this same blot with an antibody to total ERK1/2 reveals essentially equal loading of protein in all lanes.
To determine the importance of HGF-induced ERK activation for the increase in ATP-depleted cell adhesion, we examined the effects of the MEK inhibitor U0126 on adhesion of mIMCD-3 cells plated for 30 min. A dose-response curve demonstrated that 10 μM U0126 was required for complete inhibition of HGF-induced ERK1 and ERK2 activation (data not shown). Pretreatment with 20 μM U0126 prevented the baseline increase in adhesion seen in control cells after HGF stimulation (Fig.5 B) and completely reversed the HGF-dependent increase in adhesion of 90% ATP-depleted cells. Thus the increase in mIMCD-3 cell adhesion after HGF treatment requires activation of the ERK-signaling pathway.
Acute renal failure, whether caused by ischemic or toxic injury, can result in exfoliated renal tubular cells that are morphologically intact and still viable, ultimately resulting in apoptosis of these cells (35). HGF and c-met mRNA levels increase significantly in animals after acute ischemic or toxic injury (16, 18), and the administration of exogenous HGF diminishes tubular injury and accelerates renal regeneration in animal models of acute renal failure (21,29). It has recently been demonstrated that carcinoma cells stimulated with HGF will more rapidly adhere to either plastic dishes or dishes coated with various extracellular matrix components, including type 1 collagen, laminin, and fibronectin (1,48), suggesting that the HGF/c-met signaling system can regulate cell-matrix adhesion. We therefore examined the possibility that HGF can regulate adhesion in normal and ischemic renal tubular epithelial cells.
Cell adhesion is a complex, multistage process in which suspended, rounded cells initially engage matrix and form focal contacts via a divalent cation-dependent interaction with cell surface proteins known as integrins (14). The rate of engagement of the integrin with matrix is dependent on the affinity state of the extracellular domain of the integrin, a process that is not fully understood but believed to be dependent on the interaction of the cytoplasmic domain with integrin-binding proteins such as talin, α-actinin, and focal adhesion kinase (FAK). Presumably, a conformational change in the extracellular domain of the integrin then allows binding to the matrix ligand, such as the classic RGD motif in fibronectin (19). The focal contact formed by the αβ-integrin heterodimer binding to matrix matures to form a focal adhesion, and clusters of these focal adhesion complexes become major sites of actin-filament attachment and trigger the process of cell spreading. Cell spreading provides more surface area for focal contacts, resulting in progressive focal adhesion complex formation and clustering and eventually a tightly adherent cell.
The mechanism by which HGF induces increased adhesion is not well understood. In the present experiments, HGF had no effect on adhesion to poly-l-lysine [a positively charged substrate that mediates cell attachment via charge interactions with anionic proteins on the cell surface rather than by integrin engagement (40,45)] but did increase adhesion to the integrin-specific substrates fibronectin and type 1 collagen. This result suggests that HGF induces adhesion in an integrin-dependent manner in renal epithelial cells and is in agreement with two studies of carcinoma cells that have found that increased adhesion to laminin after HGF stimulation was prevented by neutralizing antibodies to α2/β1-, α3/β1-, and α6/β4-integrins (1, 48). Because HGF is known to cause “scattering” (32), a process in which epithelial cells flatten, lose their cell-cell contacts and migrate away from each other, it is possible that the effect of HGF on cell adhesion is mediated by more rapid flattening and spreading of the cell, allowing more surface area for contact with the substrate and more rapid formation of focal adhesion complexes. In agreement with this, a time course of adhesion reveals that HGF has no effect on initial attachment of mIMCD-3 cells at 5 min (suggesting that HGF does not alter the affinity state of unengaged integrins) but does increase adhesion at 30 and 60 min (consistent with an effect on focal adhesion maturation and cell spreading after initial attachment).
In an attempt to determine the potential effects of in vivo ischemia on cell adhesion, we utilized the classic in vitro model of ATP depletion. By removing cells from glucose and exposing them to the combination of the mitochondrial inhibitor oligomycin A and the glycolytic inhibitor 2-deoxyglucose, graded intracellular ATP depletion can be achieved (6, 28, 42, 43). In an effort to distinguish sublethal and lethal injury, conditions were defined whereby cells are substantially ATP depleted for 30 min but recover if ATP levels are restored. Based on the results of Venkatachalam et al. (49) and Lieberthal and coworkers (24) in proximal tubular cells, we predicted that transient 90–99% ATP depletion would be sublethal, whereas sustained 99–100% ATP depletion should be lethal to our renal tubular cells. In agreement with this, mIMCD-3 cells exposed to 30 min of 25 nM oligomycin and 100 μM 2-deoxyglucose showed ATP levels between 10 and 13% of baseline and were fully recovered when glucose was added for 1 h, whereas ATP levels in cells treated with 5 μM oligomycin and 10 mM 2-deoxyglucose for the same time declined to 1% of baseline values and recovered only to 20% of baseline values over the ensuing 60 min. Despite this slow recovery of ATP levels in the 99% ATP-depleted cells, cell survival (as judged by trypan blue exclusion) was not significantly different in transiently 90% ATP-depleted and 99% ATP-depleted cells (1.7 ± 0.6 dead and 2.7 ± 0.6% dead, respectively). In contrast, sustained 99% ATP depletion for 2 h resulted in >30% trypan blue-positive cells.
To examine the effects of ATP depletion on cell substrate interactions, we explored both the adhesion of suspended cells and the release of adherent cells. In both cases, 90% ATP depletion resulted in a significant loss of adhesion that was totally reversible after stimulation with HGF, despite no affect of HGF on cellular ATP levels. In contrast, 99% ATP depletion resulted in a more severe loss of adhesion that was only modestly affected by treatment with HGF. In a similar approach of examining cyanide-mediated ATP depletion in MPT cells, Kroshian et al. (22) found that 94% ATP depletion for 1 h resulted in a loss of 66% of the previously adherent cells and that this could be reversed by allowing ATP levels to recover for 30 min. They demonstrate disruption of the actin cytoskeleton after ATP depletion and suggest that regulation of the actin cytoskeleton may be critical for the process of adhesion in this setting. Indeed, work by Raman and Atkinson (38) confirms the disruption of actin stress fibers with ATP depletion and has demonstrated that expression of constitutively active Rho can partially reverse this effect (38). Interestingly, full reversion of the ATP-depleted cell phenotype to normal did not occur with expression of constitutively active Rho because the dispersion of vinculin (which forms the site for anchoring stress fibers to focal adhesions) and paxillin from sites of focal adhesions was not reversed.
Examination of the morphology of mIMCD-3 cells revealed that whereas control cells had extended processes and spread out 2.5 h postplating, the majority of ATP-depleted cells remained rounded. In the setting of stimulation with HGF, this failure to extend processes and flatten was partially reversed in 90% ATP-depleted cells but only minimally altered after HGF treatment in severely ATP-depleted cells. This result, coupled with the HGF-mediated increase in adhesion to the integrin substrates fibronectin and type 1 collagen, suggested to us that HGF increased adhesion by regulating integrin-dependent cell spreading and that stimulation with HGF could overcome a loss of integrin-dependent cell spreading in the setting of moderate ATP depletion.
After the binding of HGF to c-met, multiple signaling pathways are activated, including the phosphatidylinositol 3-kinase, phospholipase C-γ, src, and growth factor receptor-bound protein 2 (grb2) (4,7). The association with grb2 mediates activation of the ERK1 and ERK2 members of the MAPK signaling pathway as well as a unique association with GAB1 and its membrane localization (27,51). We have demonstrated that activated ERK2 can bind and phosphorylate GAB1 at the membrane after HGF stimulation (41), and active ERK has been recently demonstrated to associate with focal adhesions at the sites of cell attachment after src activation (11). We therefore examined the possibility that ERK activation and subsequent localization to focal adhesions at the cell membrane may be an important mediator of HGF-induced cell adhesion.
To inhibit ERK activation, we utilized the selective MEK inhibitor U0126 (10). A dose-response curve in mIMCD-3 cells revealed that 20 μM U0126 caused complete inhibition of HGF-induced ERK1 and ERK2 activation (20) but had no effect on tyrosine phosphorylation of c-met or GAB1 or on activation of the PI 3-kinase after stimulation with HGF (Yu C, unpublished observations), and this dose was therefore chosen for the present experiments. In the setting of inhibition of ERK activation, HGF failed to cause an increase in adhesion in either control or ATP-depleted cells, strongly arguing that the increase in adhesion after stimulation with HGF requires activation of the MAPK-signaling pathway. In agreement with this, HGF failed to activate ERK 1 or 2 in 99% ATP-depleted cells, a setting where HGF also failed to reverse the loss of cell adhesion. The substrate for the MAPK pathway in HGF-induced cell adhesion has yet to be identified. Recently, stimulation with HGF in breast carcinoma cells and hepatocytes has been found to cause tyrosine phosphorylation of FAK with an increased linkage between integrin complexes and the actin cytoskeleton (1, 33). Whether FAK phosphorylation is regulated by ERK activation in mIMCD-3 cells is as yet unknown.
In summary, HGF and its receptor have been found to be upregulated in the kidney after ischemic renal injury, and the administration of HGF diminishes the degree of tubular epithelial cell loss and/or promotes tubular repair. By using an in vitro ATP depletion model to partially mimic the effect of ischemia, we demonstrate the loss of matrix attachment by otherwise viable epithelial cells, and make the novel observation that this effect is completely reversed by HGF in moderately ATP-depleted cells. The ability of HGF to reverse the ATP-depletion-induced cell detachment is dependent on activation of the ERK-signaling pathway. We therefore propose that at least part of the mechanism by which HGF decreases injury after renal ischemia involves the promotion of sustained adhesion of sublethally injured cells that would otherwise detach and undergo apoptosis.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-52897 to L. G. Cantley.
Address for reprint requests and other correspondence: Z.-X. Liu, Section of Nephrology, Yale Univ. School of Medicine, 333 Cedar St., LMP 2093, New Haven, CT 06520 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society